Some people can’t be bothered to read the analog face of a traditional clock. Some people cannot stand the low frequency “hum” of mains current. If you are in either of those categories, you probably don’t want to make [Christian]’s handsome and well-documented electromechanical CMOS clock.
As you might guess from the name, the clock uses CMOS logic, based around a 12 bit counter, to provide the divider circuits 24 (daily) and 60 (minutes and seconds). Specifically, the circuits are based around a CD4040 twelve-bit adder. Those signals go through DAC circuits based around DAC0808 chips to drive some very nice coil meters for hours and minutes in lieu of the traditional clock face. Taking the time to make a CMOS clock circuit from adder chips is respectable enough in this era of instant-gratification through micro-controllers, and we dig the blinkenlights built into the circuits, but it’s what is being added that is where things get really interesting.
[Christian] had the bright idea that a stepper motor could be driven via the mains, simply by using a capacitor to offset the waveforms on the coils by 90 degrees. With a 200-step stepper motor, [Christian] gets one revolution per second out of the 50 Hz grid; this generates the seconds signal for his CMOS chips by the simple expedient of a 3D printed arm and a light barrier. Once per second, the light is interrupted by the spinning arm, creating a pulse for the clock circuits to add up. Check it out in action in the demo video below.
This project also seems to have the distinction of being the first project submitted to our One Hertz Challenge. It’s not just for clocks, but keep an eye on your clock because entries are only open until 9:00 AM Pacific time on August 19th.
I built a BCD LED clock years ago. It derived it time keeping from the supposed 60 cycle mains frequency. I say supposed because for some reason, our mains line was no where NEAR 60 cycles. It ran closer to 50 than 60 most of the time. I even built a mains frequency counter to figure out why my clock kept losing time. Finally, when the power company was no help at all, I had to build a 60 Hz generator to inject into the clock or else it would have been useless… 🤷♂️
The grid time can vary from real time by quite a bit. IIRC The frequency rises or lowers when there’s a supply and demand mismatch. It’s supposed to average out over a day, but – according to a friend who worked in the industry – in reality it doesn’t, and they reset the counter when it drifts too far.
It may drift (because many utilities don’t try and manage it for timekeeping purposes anymore) but in north america at least the typical variation might be 59.5-60.5hz at the absolute worst. If it’s down to 50hz they’re either running on a generator in alaska or somewhere with very poor power quality.
Texas has pretty poor timing from what I saw 20 plus years ago as they do their own thing
I’m in a small rural Alabama town. It wasn’t a generator. It was direct from the Alabama Power line… Might could explain why I had so much business back then. (I was a consumer electronic technician at the time)
This is like the Hammond electric clock. Legend has it Lawrence Hammond gifted a clock to the foreman at every power plant to motivate them to keep the supply at exactly 60hz so their new clock would keep good time.
I’d be keeping it SLIGHTLY above 60 so I could go home early LOL
CD4040 is a 12-bit counter not a 12-bit adder.
Don’t the North American power grids have to stay consistently close to 60hz in order to power share and sell power to one another?
The North American Continent operates several grids. There is the Eastern Interconnect, The Western Interconnect, and then there is ERCOT (in Texas). And there are smaller independent grids in less populated areas such as Alaska and Canada’s Northern Territories. These grids need to stay in Phase within the grid itself, but not exactly on frequency. During high load situations, the frequency can sag a few tens of mHz (yes, that’s millihertz), and as load drops off suddenly, it can surge a few mHz. These surges and sags are monitored with GPS synchronized devices called synchrophasers. Basically they measure how far forward or backward the grid is sliding. Over a transmission line with a significant load, you can actually measure a phase “twist.”
So it’s not always precisely on 60.000 Hz. Anomalies happen and generators do slow down and speed up depending on what those anomalies are. For example, if a transmission line phase drops to ground on a transmission stub toward a load, and the substations trip, the path from generation to load usually speeds up just a little bit. The rest of the grid will stabilize that load and while the frequency may wander about for a few mHz, the grid will survive.
If a transmission stub includes some generation, the disconnection of that stub is referred to as Islanding. There may be sufficient generation capacity to handle the stub load, or they may have to curtail some load with rolling blackouts. Before an island is reconnected to the rest of the grid, it must be in phase and on frequency.
Generally the grid is design so that any one transmission or generation site can disconnect and things will continue to run. This is referred to as N-1 conditions. There are State Estimators and Contingency Analyzer software systems for ensuring that there will always be sufficient capacity so that we don’t run things too close to the edge.
Frequency is dependant on the mismatch better demand and generation. This mismatch is carefully managed by system operators, and if it increases beyond a tiny margin it shows a system under duress. Witness what happened in the Iberian peninsula. There will be slight changes in frequency throughout the day, and here in the UK there is still the requirement of of the average over 24 hours to be exactly 50hz. In the control room of my local distribution company there is a display of frequency, NTP time and synchronous time. The latter can drift by several seconds throughout the day.
I’m in the UK and can confirm what Andrew says: the spot frequency can be slightly above or below 50Hz as the load is managed, but the average is definitely exactly 50Hz. I’ve been able to confirm this as I have some synchronous motor electric clocks, which don’t, and never have, shown any cumulative error.